JP3221594B2 - Distortion correction circuit for linearizing electronic and optical signals - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distortion correction circuit for linearizing the output from an amplitude modulation transmitting device, such as a semiconductor laser, whose output is distorted with respect to the input due to inherent non-linearity.

[0002]

2. Description of the Related Art The method of directly modulating the analog intensity of a light emitting diode (LED) or a semiconductor laser by an electric signal is the simplest method for transmitting analog signals such as audio and video signals through an optical fiber. While such analog schemes have the advantage of using a much narrower bandwidth than digital pulse code modulation or analog or pulse frequency modulation, amplitude modulation has the problems of noise and non-linearities of the light source. U.

[0003] Distortions inherent in analog transmitting devices prevent the conversion of a linearly modulated signal into an optical signal that is directly proportional thereto, and may cause signal distortion. This phenomenon is particularly detrimental for multi-channel image transmissions where good linearity is required to prevent interference between channels. Highly linearized analog optics are widely used in commercial TV transmission, CATV, interactive TV, and video telephony.

[0004] Linearization of nonlinear optical transmission devices and the like has already been studied, but the results of the research have not yet reached the level of practical use. That is, in many cases, the bandwidth is too wide to be practical. Also, the feed-forward method requires complicated system components such as a light power combiner and multiple light sources, and the quasi-optical feed-forward method is also problematic in terms of complexity. I have to.

[0005] One known method for reducing the distortion inherent in nonlinear devices is the inverse distortion application method. In this method, the modulated signal is combined with a signal of equal amplitude but opposite sign inherent to the nonlinear device. As the nonlinear device modulates this composite signal, the inherent distortion of the device is canceled by the inverse distortion of the composite signal, and only the linear portion of the source signal is transmitted. This inverse distortion signal often takes the form of addition and subtraction of the input fundamental frequencies, because such intermodulation products constitute the richest sources of distortion in analog signal transmission. For example, in an AM signal distribution for a cable TV, there are as many as 40 frequencies in a specific band, and there are abundant opportunities for forming secondary and tertiary intermodulation products of these frequencies.

[0006] In a method of applying reverse distortion to a current, an input signal is divided into two or more paths (electrical paths), and reverse distortion is generated in one or a plurality of paths corresponding to the intrinsic distortion of a nonlinear transmission device. This inverse distortion is recombined with the input signal to offset the effects of the inherent distortion of the nonlinear device.

[0007] The attenuation can be used to match the magnitude of the inverse distortion with the magnitude of the device's intrinsic distortion characteristics before the signal is recombined and transmitted to the non-linear device for modulation. However, this method has a problem in accuracy because the amplitude and phase distortion characteristics of the nonlinear device often fluctuate according to the frequency of the modulation signal.

[0008]

None of the known methods can correct for frequency dependent non-linearities.
As far as systems and signals with relatively low bandwidth are concerned, it is often not a problem to correct for the frequency dependence of the distortion. However, a serious problem arises when a TV signal is converted into an optical signal and transmitted by wire. That is, a wired TV
This type of signal can have more than 40 input frequencies, all of which require good quality amplitude modulated signals. Transmitting devices for such signals require a very high degree of linearity. The bandwidth used for TV transmissions and the like becomes considerably large, and currently available reverse distortion techniques are insufficient to cover a wide band without the desired distortion. Therefore, it is desirable to correct high-order intermodulation products over a wide frequency range. The present invention aims to overcome these problems in the prior art.

[0009]

SUMMARY OF THE INVENTION In a preferred embodiment of the present invention, a distortion correction circuit for reducing distortion associated with the transmission of an analog signal divides an input modulated signal into three paths. That is, a primary path, one even-order secondary path, and one odd-order secondary path. An inverse distortion amplifier on the even-order secondary path forms a secondary intermodulation distortion product of the input signal.
The inverse distortion amplifier provided in the odd-order secondary path is used to control the 3rd input signal.
Form the second order intermodulation distortion product. The inverse distortion formed in each secondary path is adjusted so that the amplitude is substantially equal to the intrinsic distortion of the nonlinear modulation device to which the signal is supplied, and the sign is opposite. The inverse distortion signal is amplitude and phase adjusted to match the frequency dependence of the distortion due to the nonlinear device. The phases of the signals are synchronized by delay or phase adjustment elements provided in each path. Then primary and 2
A single modulated signal containing intermodulation product distortion is formed by recombining the signals in the next path. The phases of the even-order secondary paths are adjusted at both high and intermediate frequencies within the frequency range of the circuit. Thus, the distortion correction circuit of the present invention can effectively linearize the transmission of the modulated signal by canceling out the inherent distortion of the nonlinear transmission device over a wide frequency range.

[0010]

FIG. 4 illustrates the concept of inverse distortion in an abstract manner.
It is. The input signal Y ₀ is input to the distortion correction circuit 40. The distortion correction circuit 40 has a nonlinear transfer function whose transfer function deviates from linearity by the same amount in the opposite direction to the known nonlinear transmission device 41. The signal Y ₁ from the distortion correction circuit 40 is a composite signal of the input source signal Y ₀ and the inverse distortion due to the nonlinear transfer function of the correction circuit 40. The signal Y ₁ is provided to a non-linear transmitting device and after being modulated by the transmitting device, the inherent distortion of the transmitting device is inversely proportional to the inverse distortion of the signal Y ₁ and is thereby canceled out, resulting in a substantially linear signal Y 1 Appears as ₂ .

As is apparent from FIG. 1, an input source signal 12 is supplied to a directional coupler 10 and divided into a primary path 13 and a secondary path 14. Typically, the power of the primary path signal is much greater than the power of the secondary path signal.
For example, such a result can be obtained by using an 11 decibel (dB) directional coupler.

The secondary path is the distortion generator 15, amplitude adjustment block 17, "tilt" or frequency adjustment block 1.
9 and a phase fine adjustment block 21. Note that the order of these elements along the secondary path can be changed without departing from the purpose of the invention.

In one embodiment of the present invention, the secondary path signal is first input to a distortion generator. The output of the distortion generator is the intermodulation distortion of the input frequency. Second or higher order strain can be generated. Ideally, it is preferable to suppress the fundamental frequency by canceling, filtering, or the like in the distortion generator. The phase of the intermodulation product thus formed is opposite to the input signal. This inversion can be performed in the strain generator or in a separate inverter element (not shown).

The magnitude of the strain output from the strain generator matches the foreseeable natural strain magnitude of the transmitting device (not shown in FIG. 1) that receives the output signal 25. This matching action is performed in an amplitude adjustment block 17, which can be done manually, for example using a variable attenuator, or dynamically using an automatic gain control element. Accordingly, the output of the amplitude adjustment block 17 is the intermodulation product of a small portion of the input signal, and the magnitude is substantially equal to the intrinsic distortion of the nonlinear transmission device 41 that receives the output signal 25 of the distortion correction circuit 40, and the sign is opposite. It is. This output or inverse distortion signal effectively reduces the frequency dependent distortion component of the nonlinear device 41.

The generation of the inverse distortion signal in the secondary path is 1
Usually, there is a delay for the next path. Before the primary and secondary paths are recombined, the relative phase of the primary path signal to the phase of the secondary path signal is adjusted so that the intrinsic distortion of the non-linear device is effectively canceled. This phase matching is performed in the primary path by external delay means 23 that receives the primary path portion of the signal 13 divided by the directional coupler 10. The amount of delay can be adjusted manually or automatically. As the delay means, for example, there is a simple means such as a transmission line whose length is set so as to introduce an appropriate delay.

Examples of the transmitting device include a semiconductor laser or an LED modulated by an output signal. The intrinsic distortion of such devices is frequency dependent. In general, the higher the frequency, the greater the distortion.

The output of the amplitude adjustment block is provided to a frequency adjustment or "tilt" adjustment block 19 to correct for frequency dependent distortion of the nonlinear transmission device. Tilt adjustment is performed by means such as a variable filter that increases the amplitude of the high frequency distortion as "up-tilt" and reduces it as "down-tilt". This adjustment can be performed manually or automatically similarly to the amplitude adjustment.
In the tilt adjustment, by adjusting the ratio of the high-frequency distortion product to the low-frequency distortion product to be passed, the inverse distortion signal can be adjusted as accurately as possible to the inherent distortion characteristic of the nonlinear device.

Normally, amplitude adjustment is performed to correct distortion appearing at the low-frequency end of the band, and frequency adjustment is performed as up-tilt to correct distortion appearing at the high-frequency end of the band. Note that the same result can be obtained by adjusting the amplitude at the high frequency end and up-tilt or down-tilt at the low frequency end as appropriate attenuation or amplification of the signal.

The fine adjustment block 21 in the secondary path is 2
This makes it possible to more accurately set the relative phase between the distortion generated in the next pass and the intrinsic distortion of the nonlinear device. This adjustment may be performed manually and in accordance with the frequency, similarly to the amplitude adjustment. Experience shows that manual adjustment of amplitude, frequency and phase can be completed in less than one minute. In that case, appropriate adjustment may be made while observing the output distortion of the nonlinear device. The purpose of this adjustment is to make the final strain as small as possible. If the inverse distortion signal has the same magnitude as the intrinsic distortion of the nonlinear device, and the inverse distortion has exactly 180 ° of phase shift relationship with the distortion, the optimal adjustment has been made.

It is important to make phase adjustments for device distortion. Since a delay has already been introduced, the inverse distortion is exactly in phase (or 180 ° out of phase) with the primary path signal. No further adjustment is required for some purposes, but it is not suitable, for example, for laser TV bandwidth modulation.

After setting the relative phases of the signals of the primary and secondary paths, the directional output coupler 11 recombines the two signals. This combined output signal 25, which contains the inverse distortion component from the secondary path, is output to a non-linear transmission device to modulate the signal.

As an example, the inverse distortion forming or distortion amplifying block 15 is illustrated in detail in FIG. Input signal 1
The part passing through the secondary path of 4 is a 180 ° splitter 3
0, the splitter 30 converts this signal into a first path 38 and a second path 39 of equal magnitude and opposite sign.
Divided into If amplified or attenuated later, the split signals need not be the same magnitude.

The first path is the fundamental frequency 2 of the input signal 14.
The signal is input to a first amplifier 32 that forms an intermodulation product of the next order or higher. A second path, which carries a signal opposite the sign of the first path signal, enters a second amplifier, which has an even-order intermodulation having the same sign as the intermodulation product output from the first amplifier 32. And the odd-order intermodulation products of opposite sign. The signal is additively combined by a zero (0) degree combiner 34 that substantially eliminates the fundamental frequency and odd order intermodulation products, leaving an even order intermodulation product component in the output signal 37. Ideally, it is desirable that a pure even-order component of the second order or higher of the intermodulation distortion be formed in this process.

The first and second amplifiers 32 and 33 are adjusted so that the odd-order intermodulation product components are not completely eliminated.
This adjustment can be achieved by changing the bias current to the amplifier, and changing the bias current has little effect on the fundamental frequency gain. The bias current of the first amplifier 32 increases and the second amplifier 3
As the bias current of 3 drops, both amplifiers become unbalanced, causing a difference in the magnitude of the intermodulation product formed by both amplifiers. Therefore, the odd-order intermodulation products do not cancel each other.

By utilizing the unbalanced state of this distortion circuit called a push-push amplifier, it is possible to generate intermodulation distortion in all dimensions for the purpose of distortion correction. To suppress the fundamental frequency (not shown)
This can be done by specially designing the amplifier or by providing filter means behind or in series with the amplifier or integral with the amplifier. Preferably, the bias currents of both amplifiers 32, 33 are adjusted in the same and opposite directions so that the imbalance only affects the odd-order intermodulation products and the magnitude does not change substantially while the even-order intermodulation products remain in a balanced state. adjust.

One embodiment of the distortion correction circuit of the present invention is the third embodiment.
Shown in the figure. Secondary path signal 1 from signal division coupler 10
4 is first attenuated by the adjustable attenuators R ₁ and R ₃ to secure a constant signal level. If the signal is too small, enough reverse distortion to correct the distortion of the transmitting device cannot be obtained, and if it is too large, the correction circuit will be overloaded,
This results in unacceptably large strains.

The attenuated signal is split by a 180 ° splitter 30 and capacitively coupled to first and second amplifiers 32,33. The required third-order or higher intermodulation product is obtained by adjusting the bias of both amplifiers, and the recombined signal is attenuated by the amplitude adjustment 17, for example, to 50 MHz.
Obtain the required amount of low frequency distortion on the order of z. The high frequency end of the band is then checked and the frequency filter 19 is checked until the distortion matches the intrinsic distortion of the transmitting device at this high frequency.
To adjust. This has little effect on the reverse distortion at the low-frequency end of the band, and the amplitude is tilted according to the frequency around the low-frequency end of the band.

The phase of the primary path signal is adjusted by adjusting the delay 23 at the high frequency end of the band. This also has little effect at the low frequency end of the band. Finally, by using the phase adjustment 21, the phase of the reverse distortion generated in the secondary path for correcting the phase distortion of the nonlinear device is more accurately adjusted. The adjustment sequence can be repeated as needed to more accurately match the intrinsic distortion of the transmitting device. In many cases, the adjustment of the initial attenuator and the reverse distortion amplifier bias is not necessary and can be left in the preset state. Only three adjustments are needed: amplitude, tilt and phase. The basic delay in the primary path may be constant for a given secondary path.

The primary path signal 13 and the secondary path signal 14 are recombined by the directional coupler 11, and the output signal 25 given the reverse distortion is supplied to a laser 42 or the like for modulation.

In the above embodiment, there is one secondary signal path including the distortion generator. However, if necessary, a third "secondary" path 46 is provided as shown in FIG. That is, one secondary path 47 forms a secondary cancellation signal, and the other secondary path 46
Form a third order cancellation signal. In FIG. 5, the reference numbers of the secondary paths are the same as those of FIG.
The series is numbered 200 and 200. In each of these paths, the frequency dependence of the amplitude and phase 119,
Adjustments to 219 may be made. In that case, it is preferable to finely adjust the phases 121 and 221 in each of the secondary paths. When two or more secondary paths are used to form higher-order distortion, there is no interaction between the two paths, so that adjustment of amplitude, tilt, and phase may be started from either path.

The above embodiment of the distortion correction circuit uses a single 2
Next path, which is configured to be used over a frequency range of 50 to 300 MHz, which reduces distortion of the optical device from 50 to about 450 MHz.
Has been found to be sufficient to correct over a range of frequencies. However, there are TV signals that use the frequency band up to 860 MHz, so it is desirable to cover a wider range of frequencies. The distortion at the high frequency end of such a band is significant and the simple circuits shown in FIGS. 2 and 3 are not sufficient for a wide frequency range. Accordingly, an improved distortion correction circuit covering a wider frequency range is shown in FIG.

As shown in this embodiment, the radio frequency (R
F) The input signal is provided to combiner 50, where the signal is split, with the main part of the signal being transmitted along primary path 51 and the remaining minor part being transmitted to the secondary path. As in the previously described embodiments, the primary path includes a delay 52, which more or less matches the delay inherent in the multiple secondary paths described below.

The signal in the primary path is connected to another combiner 53
Recombined with the signal in the secondary path. This recombined signal has a distortion correction that is opposite to the inherent distortion of an output device (not shown) such as a laser. 75 between the distortion compensation circuit and the modulating working device
It is desirable to use an impedance matching device 54, such as a ohm to 25 ohm transformer.

The weak signal from the input coupler 50 is split by a splitter 55 into an even-order secondary path 56 and an odd-order secondary path 57. The even-order secondary path is provided with means for forming a secondary intermodulation product that is adjusted to be equal to and opposite to the distortion of the output device. The odd-order secondary path is provided with means for forming a third-order intermodulation product which is equivalent to, and opposite to, the distortion of the output device. 2
The second and third order distortion correction signals are combined by a combiner 58, and further combined by a combiner 53 with the signal of the primary path 51.

The even-order secondary path is an initially adjustable attenuator 5.
9 inclusive. This attenuator and the other adjustable attenuators shown in FIG. 6 are substantially the same as the attenuators R ₁ , R ₂ , R ₃ shown in FIG. The adjustable attenuator reduces the signal strength to prevent overloading the push-push secondary distortion generator 61 to which the signal is applied.

Suitable push-push strain amplifier 6
1 is shown in FIG. The amplifier includes a 180 ° splitter 62, which splits the signal into two parallel paths of equal magnitude and opposite sign. Each of the divided signals is applied to the same amplifier 63. The output signals from these amplifiers are recombined and these are
° Out of phase. As a result, the fundamental frequencies cancel out as odd-order intermodulation products. The second higher even-order intermodulation products are left intact.

The second-order product from the push-push amplifier is applied to an all-pass delay equalizer 64 which is a circuit shown in FIG. The all-pass filter 64 splits the signal into portions with equal and opposite signs.
including. Each signal from the splitter is applied to an inductor 67. The input to one inductor is connected by a capacitor 68 to the output of the other inductor. Output signal is 180
° Recombined by coupler 69. This all-pass filter has a flat magnitude response but provides an appropriate phase delay as a function of frequency.

As described below, this all-pass delay equalizer is used to adjust the phase of the second order intermodulation product within the intermediate frequency range. In practice, some "plug-in" equalizers are used that provide an acceptable correction rather than a continuously adjustable one. In this all-pass delay equalizer, the amount of correction in the intermediate frequency range is increased by increasing the induction and capacitance. In this device, the square root of the induction for the capacitance is kept constant at about 50 ohms. If desired, a delay equalizer utilizing an operational amplifier may be used in place of the LC device. But LC
The device is cheap and convenient. This is because only a very small number of devices of different characteristics need to be provided for quick adjustment.

The output from the delay equalizer 64 goes to a switchable RF inverter 65 and another adjustable attenuator 71 and from there to a buffer amplifier 72. This inverter is used because the second order inverse distortion required by a particular output device is positive or negative with respect to the fundamental frequency. The amplifier output is applied to a coarse delay adjuster 73, which is typically the length of a plug-in coaxial cable. Fine adjustment of the delay is provided by the variable capacitance 74.

The output of the delay adjustment is applied to the amplitude tilt circuit 76 via the buffer amplifier 77. The output from the amplitude tilt circuit goes to a combiner 58 and is combined with a signal from an odd-order secondary path. The delay adjustment and the amplitude adjustment are the same as in the above-described embodiment.

FIG. 10 shows a suitable amplitude tilt circuit. If the signal line is grounded via the variable resistor 78, the capacitor 79 and the variable inductor 81, a tilt can be obtained. In practice, it is easier to change the tilt by inserting any of several inductors of different values instead of one changing inductor. Also, various capacitors can be inserted for tilt adjustment.

The capacitor 79 is selected for adjusting the phase of a signal near the low frequency end of the frequency of the reverse distortion circuit.
This can be changed without significant changes in the amplitude at the low frequency end of the frequency range. By properly selecting the induction and resistance values, the amplitude of the strain as a function of frequency can be set to match the device being linearized.
Generally speaking, the amplitude of the high frequency end is adjusted and then the resistance of the tilt circuit is changed to change the amplitude of the low frequency end within the frequency range of the device. A change in the induction in the tilt circuit changes the amplitude in the intermediate frequency range.

The signal on the odd secondary path 57 goes to an adjustable attenuator 82 and from there to a push-pull tertiary distortion amplifier 83. FIG. 8 shows a suitable push-pull circuit. In this circuit, the RF signal is split by a 180 ° splitter 84 and the resulting signals are applied to the same amplifier 86. These amplifier outputs are recombined in a 180 ° combiner 87. Because of the phase inversion, the circuit that produces the odd-order intermodulation products includes the third-order frequency and the fundamental frequency.

The magnitude of the required third-order distortion correction is usually quite low (eg, less than 60 dB), and the magnitude of the fundamental frequency appearing in the odd-order secondary path is smaller than the gravity in the primary path. Turned out to be insignificant. Therefore,
No special measures are usually required to further suppress the fundamental frequency in the odd-order path.

By comparison, the secondary signal is typically about 45 dB below the fundamental frequency when recombined with the signal on the primary path. The magnitude of the fourth and higher order intermodulation products is very low, typically less than 75 dB below the fundamental frequency. Of course, these values depend on the device to be corrected.

The output of the third-order distortion amplifier is applied to an amplitude tilt adjustment circuit 88 similar to that described above. The output of this tilt adjustment circuit goes to a switchable RF inverter 89 which is used to change the polarity of the signal. The use of this inverter requires three
This is because the second-order inverse distortion is positive or negative with respect to the fundamental frequency. The odd secondary path also includes a coarse delay adjustment 91,
This adjusts the delay of the third order intermodulation product independently of the second order intermodulation product before the two inverse distortion signals are combined at combiner 58.

As noted above, the order of the circuit components along the secondary path is not critical. This is shown in FIG. 6 at the position of the variable capacitor 92 for fine adjustment of delay, with some circuit elements removed from the coarse delay adjustment. This is because it happens to be a convenient place to make fine adjustments in the example device. Other circuit elements can be arranged in different orders. Another example is an inverter that changes the sign of the third-order intermodulation product. this is,
For example, this can be achieved by switching the input line to a push-pull distortion amplifier. Other changes are obvious.

The improved inverse distortion circuit provides good correction of distortion over a wide frequency range due to several other features. One is the separate odd order secondary path to form the third order intermodulation products. This allows the use of a balanced push-push inverse distortion generator on the even-order secondary path to form a second-order intermodulation product that substantially completely cancels the fundamental frequency. Thus, the distortion introduced by the subsequent amplifier in the even-order path can be minimized. Next, the all-pass delay equalizer enables phase adjustment at an intermediate frequency in the frequency range of the inverse distortion circuit, in addition to phase adjustment near the high-frequency end in the frequency range of the inverse distortion circuit.

Adjustment of the improved inverse distortion circuit is generally similar to that of the simpler circuit, but has more steps and takes a little longer. One skilled in the art can set the inverse strain generator in a few minutes. Adjust each inverse distortion generator to match the distortion inherent in the device being modulated. This is because lasers and the like have their own distortion characteristics. Some devices readjust the inverse distortion generator to compensate for new distortions when changing lasers. This adjustment is performed by applying a known signal and observing the distortion of the output device. These adjustments are made to reduce the distortion seen at the output.

The general order of adjustment begins by first balancing the push-push amplifier, forming only second-order distortion, and suppressing the fundamental frequency. Next, the push-pull distortion amplifier is balanced to form only the third-order inverse distortion. At this point, it is desirable to determine the relative poles of the second-order and third-order distortions with respect to the fundamental frequency and set each RF inverter.

Usually, the second-order inverse distortion is adjusted first. Adjustable attenuators and delay adjustments are used to cancel relatively high frequency distortion near the high frequency end of the frequency range of interest. This distortion usually vanishes at some frequency below the end of the frequency range. The amplitude tilt adjustment is then used to eliminate distortion near the low end of the frequency range. Check the distortion near the upper end of the frequency range, and if necessary, repeat the attenuator, delay adjustment and tilt adjustment to further reduce the distortion at the high frequency end and low frequency end of the frequency range.

Next, the all-pass delay equalizer is adjusted to minimize the phase difference at the intermediate frequency within the frequency range of the inverse distortion circuit. If the tilt adjustment is appropriate, perform the tilt adjustment again to optimize the mid-range amplitude. Check the distortion near the high and low frequency ends of the frequency range, and repeat the attenuation, delay adjustment and tilt adjustment if the above various adjustments are necessary.

The result of this three-point adjustment can be seen, for example, from FIG. 11 which graphically illustrates the distortion d as a function of the frequency f for the frequency range 50 to 860 MHz.
The curve shown by the solid line indicates an arbitrary level of amplitude or phase distortion of the non-linear output device. The curve shown by the dotted line eliminates distortion at frequencies near 50 MHz, distortion at relatively high frequencies near 860 MHz, and distortion at intermediate frequencies. The dotted curve shows the amount of reverse strain subtracted from the intrinsic strain. Thus, the final distortion of the output is the difference between these two curve lines. By offsetting the three points of such distortion, the net distortion of the output is
It is evident that it is less than when only the distortion near the high and low ends of the frequency range is canceled.

After adjusting the even-order secondary path, the odd-order secondary path is adjusted. First, adjust the attenuation and delay elements to eliminate the three inverse distortions near the high frequency end of the frequency range. Next, the tilt is adjusted to eliminate distortion near the low frequency end of the frequency range. Then check the high frequency end and repeat these adjustments if necessary. These adjustments are repeated until the desired linearity of the output device is obtained. In the odd-order secondary path, adjustments at the high end and low end of the frequency range are sufficient for most purposes, and usually no adjustment is made at intermediate frequencies.

When two secondary paths are used for adjusting the third-order distortion separately from the even-order paths, the amplitude, tilt, and phase of one of the paths are adjusted first. Because the techniques here are worthwhile, in addition to compensating for intrinsic distortions such as lasers to obtain better TV signals, lasers with intrinsic distortions that are larger than acceptable distortions can also be used. There is an acceptable level of distortion for a given laser applied. Also, lasers with excessive strain are not used. However, by giving an appropriate reverse strain, an unusable laser can be used, and the yield of the manufacturing process is increased.

It may be desirable only to automatically adjust the inverse distortion circuit to compensate for changes that occur in the nonlinear device. For example, laser operation over several years may age and change the distortion of the laser so that the original reverse distortion circuit setting no longer completely minimizes the distortion of the laser output. Under these circumstances,
When intermittent recalibration is desired, connect to a recalibration device that automatically detects residual distortion in the laser output and automatically re-adjusts the inverse distortion circuit adjustment settings to minimize residual distortion. .

In some cases, it is not desirable to automatically correct the residual strain at shorter intervals in order to correct a change in strain caused by a change in temperature or the like. If this is desired, the recalibration circuit can be permanently connected to the reverse distortion circuit to intermittently detect residual distortion in the laser output and change the settings of the reverse distortion circuit to minimize residual distortion. The recalibration circuit can be operated intermittently or even continuously during use of the laser and without interfering with the normal operation of the laser. This recalibration technique is the same for the two situations, the only difference being the desired frequency of recalibration.

The recalibration system can be understood from FIG. 12 for an inverse distortion generator. The same reference numerals used in FIG. 12 as those shown in FIG. 1 have the same reference numerals as those in FIG. For example, the input signal 12 in FIG. 1 is 112 in FIG. In this inverse distortion generator, an input source signal 112 is provided to a directional coupler 110 where a primary path 113 and a secondary path 1
It is divided into fourteen. This secondary path is the distortion generator 1
5. It has an amplitude adjustment block 117, a tilt or frequency adjustment block 119, and a phase fine adjustment block 121 in series. This circuit output is used to modulate a non-linear device 90 such as a semiconductor laser.

The output of distortion generator 115 includes the intermodulation distortion of the input frequency. Form secondary or secondary and higher order strains. The fundamental frequency is suppressed. The amplitude adjustment block includes a non-linear device 90, such as a modulated semiconductor laser.
Used to adjust the amplitude of the strain to match the expected intrinsic strain of the

The modulated signal 124 in the primary path is
3, bringing the primary and secondary path signals into a coarse phase. Secondary path phase fine adjustment block 1
21 is used to bring the inverse distortion of the secondary path into the distortion and phase inherent in the nonlinear device (more precisely, out of 180 ° phase).

To adjust the frequency dependent distortion of the non-linear device, the output of the amplitude adjustment block is provided to a tilt or frequency adjustment block 119. The tilt adjustment means is a variable filter, which changes the amplitude of the distortion at high frequencies relative to the amplitude at low frequencies to more closely match the distortion characteristics inherent in the nonlinear device. The distortion signal 122 formed by these circuit elements in the secondary path is combined with the primary modulation signal 124 by the second directional coupler 111 and applied as an input signal to the non-linear device. The operation of this inverse strain generator is the same as described above.

To adapt the inverse distortion generator to changes in the distortion of the non-linear device, a small portion of the output signal (light beam in the case of a laser) is directed by a splitter 91 to a receiver 92. The receiver detects any residual distortion in the output signal from the device and generates a distortion level signal 93 to the recalibration control circuit 94.

The receiver is set to isolate frequencies at which distortion is expected to detect residual distortion at the output of the nonlinear device. As an alternative, a set of pilot tones 95 is applied to the input signal via a directional coupler or splitter 96. Then, the receiver is set to detect distortion from the input pilot signal sound. For example, if the pilot signal sound has frequencies f ₁ and f ₂
And contains Then the receiver has the frequency f ₁ + f ₂
And f ₁ −f ₂ are detected. Several such pilot tones are input at various points within the frequency range of the system to cancel distortion anywhere within the frequency range.

The recalibration control circuit 94 checks the distortion level of the output according to the algorithm, changes the setting of the inverse distortion circuit, senses the change in the distortion level, and if the change of the setting of the inverse distortion circuit is appropriate. Make that change to minimize distortion. The control circuit uses several algorithms to make adjustments at multiple set points to minimize distortion.

The method that is easiest to understand is a method that includes a sequence increment change at a set point of each parameter.
For example, first adjust the amplitude setting. First, the control circuit increases the set point (eg, voltage) in small increments and measures the change in distortion as a result of the increment change. If the distortion is increasing, reverse the incremental adjustment. That is, the control circuit performs incremental adjustment in the opposite direction. If the strain increases again, it is known that the original set point was correct, and the control circuit returns to the original set point.

On the other hand, if the distortion is reduced as a result of the incremental adjustment, the incremental adjustment at the set point is further performed, and the distortion is measured again. Make additional incremental changes as long as each increment reduces distortion. If an additional incremental change in the set point results in increased distortion, it is known that the best set point has been exceeded, and the control circuit reverses the incremental adjustment.

Once the best amplitude set point has been found, a series of similar incremental adjustments are made to the phase fine adjustment set point. This incremental adjustment continues until the best set point for phase adjustment is achieved.

Next, the tilt increment is continuously changed, and the low frequency distortion is measured to determine the best tilt adjustment setting. Amplitude adjustment, phase adjustment, and tilt adjustment are not completely independent. Therefore, when the “best” phase adjustment and tilt adjustment are performed, the amplitude adjustment is no longer the best setting. Therefore, it is desirable to make continuous incremental adjustments of the amplitude setting and these adjustments at each set point until the desired minimum distortion is obtained.

Other types of adjustment algorithms are well known and used for other purposes. For example, a more complex technique, called the gradient search method, is used to adjust multiple parameters simultaneously, which provides the best settings more quickly than the sequence increment adjustment described above.

One of these techniques is to adjust the adjustment settings from 1 to
Modulate at a low frequency as low as 10 Hz and observe the distortion modulation at this frequency. If the modulation of the distortion is in phase with the modulation of the adjustment setting, the adjustment will be too high, and if the modulation of the distortion is out of phase, the adjustment setting will be too low. If the adjustment setting is the best, zero modulation of distortion is obtained at the modulation frequency of the setting, and distortion modulation at twice the input frequency is observed instead.

In this gradient search technique, the second adjustment parameter can be modulated at the second frequency and is measured simultaneously with the first modulation. Additional set point adjustments modulate at additional frequencies. Thus, in principle, any number of set point adjustments can be made simultaneously to obtain a quick recalibration. The inverse strain generator described here typically has three or four
Only one set point adjustment needs to be made in the even-order secondary pass. Three or four set point adjustments may be made in the third pass.

Other automatic algorithmic methods can be used for adjusting the amplitude, phase and tilt set points. Automatic adjustment of the inverse strain generator typically requires the use of resistors or capacitors that change, for example, in response to voltage. A suitable variable resistor is a pin diode.
Varactor diodes are suitable voltage variable capacitors.

The intermittent adjustment of the set point of the inverse distortion generator should be distinguished from forward feed linearization. In forward feed linearization, the sample is a continuation of the output from the non-linear device. This sample is compared with the input signal and the difference is determined as distortion. The difference is amplified as necessary and applied to a device similar to the non-linear device to form a second output signal corresponding to the distortion. The second output signal is added to the distortion output with an appropriate delay, and an appropriate phase relationship is provided between the added signals to correct the distortion. In other words, instead of reverse-distorting the input signal to the nonlinear device, an opposite distortion is applied to the output of the nonlinear device to cancel the distortion. This change is made continuously in real time distortion.

Although the above description has been directed to a TV signal for modulating a semiconductor laser and a light emitting diode, the present invention is applicable to other non-linear devices such as an amplifier.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a distortion correction circuit.

FIG. 2 is a block diagram of a push-push amplifier used in a distortion correction circuit according to a preferred embodiment of the present invention.

FIG. 3 is an explanatory diagram schematically showing a distortion correction circuit.

FIG. 4 is an explanatory diagram showing the influence of reverse distortion on a modulation signal waveform.

FIG. 5 is a block diagram of a distortion correction circuit having two or more “secondary” paths.

FIG. 6 is a block diagram of another embodiment of a distortion correction circuit effective over a wide frequency range.

FIG. 7 is a circuit diagram of a push-push amplifier.

FIG. 8 is a circuit diagram of a push-pull amplifier.

FIG. 9 is a circuit diagram of an all-passband delay equalizer.

FIG. 10 is a diagram illustrating an amplitude tilt circuit.

FIG. 11 is a graph showing the correction of the distortion of the nonlinear device.

FIG. 12 is a block diagram showing an embodiment of a distortion correction circuit having an inverse distortion generator.

[Explanation of symbols]

10, 110 Directional coupler 11, 111 Directional output coupler 12, 112 Input source signal 13, 113 Primary path (signal) 14, 114, 214 Secondary path (signal) 15, 115, 215 Strain generator 17 , 117, 217 Amplitude adjustment 19, 119, 219 Frequency adjustment 21, 121, 221 Fine phase adjustment 23 External delay means 25 Synthetic output signal 30 180 ° splitter 32 First amplifier 33 Second amplifier 34 Zero-degree combiner 37 Output signal 40 Distortion Correction circuit 41 Nonlinear transmission device 42 Laser 54 Impedance matching device 59,71 Attenuator 64 All-pass delay equalizer 65,89 Inverter 72,77 Buffer amplifier 73,91 Rough delay adjuster 76,88 Amplitude tilt circuit 83 Distortion Amplifier 90 Non-linear device 93 Level signal 4 recalibration control circuit Y ₀ input source signals Y _1, Y ₂ signals R _1, R _2, R ₃ adjustable attenuator

──────────────────────────────────────────────────の Continuation of front page (72) Inventor Howard L. Roboda United States, 90035 South Worcester Street, Los Angeles, California 1128 (72) John S. Inventor. Frames United States, 90503 California, Trance, Palos Verdes Boulevard 21513 (56) References JP-A-60-25310 (JP, A) US Patent 3,732,502 (US, A) (58) Fields studied (Int. Cl. ⁷ , DB name) H04B 3/06 H04B 1/04 H04B 10/18

Claims (13)

(57) [Claims] Dividing an input modulated signal into a primary path, an even-order secondary path, and an odd-order secondary path; forming a secondary intermodulation distortion in the even-order secondary path; Adjusting the magnitude to be equal to the magnitude of the intrinsic distortion of the non-linear modulation device but opposite in sign; and adjusting the amplitude of the intermodulation distortion of the even-order secondary path to the intrinsic frequency dependence of the modulation device. Adjusting the frequency function to match the distortion, forming an tertiary intermodulation distortion in the odd-order secondary path, and adjusting the magnitude of the intermodulation distortion to be equal to the intrinsic distortion of the nonlinear modulation device. and adjusting to also sign opposite, the amplitude of the intermodulation distortion of odd order secondary path, and adjusting as a function of frequency to match the natural frequency dependent distortion of the modulation device, the primary signal and the 2 Recombining the signals to provide an output signal having a frequency dependent intermodulation inverse distortion that cancels the distortion of the nonlinear modulation device, reducing the distortion of the amplitude modulated signal from the nonlinear modulation device. A distortion correction method for linearizing the characteristic electronic signal and optical signal. Dividing the input modulated signal into a primary path, an even-order secondary path, and an odd-order secondary path; forming a secondary intermodulation distortion in the even-order secondary path; the size, specific strain magnitude and Thus with equal nonlinear modulating device also comprising: code is adjusted to be opposite, the odd next secondary path third-order intermodulation distortion formed, the intermodulation Adjusting the magnitude of the distortion so that the magnitude of the distortion is equal to the magnitude of the intrinsic distortion of the non-linear modulation device and the sign is opposite to the magnitude of the distortion; and adjusting as a function of frequency matching dependent distortion and frequency, the amplitude of the intermodulation distortion in the even next secondary path in the low-frequency end of the frequency range, the non-linearity of the device at that frequency
And adjusting in accordance with the amplitude of the specific strain, the amplitude of the intermodulation distortion in the even next secondary path in the high frequency end of the frequency range, the non-linearity of the device at that frequency
And adjusting in accordance with the amplitude of the specific strain, the amplitude of the intermodulation distortion in the even next secondary path in the middle portion of the frequency range, the non-linearity of the device at that frequency
And adjusting in accordance with the amplitude of the specific strain, and recombining the primary signal and secondary signal, the non-linearity modulation device
Frequency-dependent intermodulation inverse distortion that cancels out
A method of linearizing an electronic signal and an optical signal , the method comprising reducing distortion of an amplitude modulation signal from a non-linear modulation device. 3. A means for dividing an input modulated signal into a primary path, an even-order secondary path, and an odd-order secondary path;
Means for forming a second order intermodulation product; and at least three input frequencies in an odd second order path.
Means for forming secondary intermodulation; delay means for reducing the relative position difference between the primary and secondary paths in the primary path; and modulation of a nonlinear device having predictable distortion characteristics. Means for combining the signals of the primary and secondary signal paths into a single signal, and intermodulation products of the two secondary paths and distortion of the non-linear device within each of the secondary paths. delay means for correcting the relative phase difference between the, there in each secondary path, the relative amplitude of the signal in that path
Adjusting means for adjusting as a frequency function , pre-distorting the intermodulation signal, and canceling the frequency-dependent distortion of the nonlinear device, wherein the distortion correcting circuit linearizes the electronic signal and the optical signal. . 4. A means for splitting the input modulated signal of the non-linear device into a primary path and a secondary path, and at least a second order intermodulation product having a relative amplitude corresponding to the amplitude of the distortion of the non-linear device. Means for generating in the next path, and means for generating an intermodulation product of the modulated signal in series, wherein the amplitude of the signal of the secondary path is adjusted as a frequency function,
Means for generating the frequency-dependent inverse distortion of the secondary path, means for adjusting the relative position of the intermodulation product and the distortion of the non-linear device, and addition and recombination of the signals of the primary path and the secondary path to obtain the nonlinearity. Means for forming a single signal consisting of a fundamental signal and a second-order intermodulation product in phase with the distortion of the non-linear device as a signal applied to the device; and a distortion remaining at the output of the non-linear device. An electronic signal comprising: means for detecting; and means for intermittently changing the determination of the amplitude adjusting means in response to the detecting means to minimize residual distortion of the output of the non-linear device. A distortion correction circuit that linearizes an optical signal. 5. A means for splitting an input modulated signal of a non-linear device into a primary path and a secondary path; Means for generating in the next path, and means for generating an intermodulation product of the modulated signal in series, wherein the amplitude of the signal of the secondary path is adjusted as a frequency function,
Means for generating the frequency-dependent inverse distortion of the secondary path; and adding and recombining the signals of the primary path and the secondary path, and applying the distortion and phase of the basic signal and the nonlinear device as signals to be applied to the nonlinear device. Means for forming a single signal consisting of matched second order intermodulation products; means for detecting residual distortion at the output of the nonlinear device; and intermittent determination of the amplitude adjustment means in response to the detection means. Intermittently determining the means for adjusting the intermodulation product and the relative phase of the non-linear device in response to the detecting means. Means for minimizing residual distortion of the output of the non-linear device by changing the electronic signal and the optical signal into linear signals. 6. The mutual transformation in the even-order secondary path.
The step of adjusting the amplitude of the tuning distortion as a frequency function
The magnitude of the intermodulation distortion in the second-order path is relatively small.
Large distortion of nonlinear modulation device at low frequency
Phase that occurs at a relatively high frequency
The magnitude of the intermodulation distortion is
Is substantially equal to the magnitude of the distortion of the linear modulation device.
The intermodulation distortion that occurs at relatively low frequencies
Adjust tilt without substantially changing the size
2. The method according to claim 1, further comprising the steps of:
Law. 7. The mutual transformation in the odd-order secondary path.
Adjusting the amplitude of the tuning distortion as a function of frequency is an odd step.
The magnitude of the intermodulation distortion in the second-order path is relatively small.
Large distortion of nonlinear modulation device at low frequency
Phase that occurs at a relatively high frequency
The magnitude of the intermodulation distortion is
Is substantially equal to the magnitude of the distortion of the linear modulation device.
The intermodulation distortion that occurs at relatively low frequencies
Adjust tilt without substantially changing the size
2. The method according to claim 1, further comprising the steps of:
Law. 8. The mutual transformation in the even-order secondary path.
The step of adjusting the amplitude of the tuning distortion as a frequency function
The magnitude of the intermodulation distortion in the second-order path is relatively small.
Large distortion of nonlinear modulation device at high frequency
Phase that occurs at a relatively low frequency
The magnitude of the intermodulation distortion is
Is substantially equal to the magnitude of the distortion of the linear modulation device.
The intermodulation distortion that occurs at higher frequencies
Adjust tilt without substantially changing the size
2. The method according to claim 1, further comprising the steps of:
Law. 9. The mutual transformation in the odd-order secondary path.
Adjusting the amplitude of the tuning distortion as a function of frequency is an odd step.
The magnitude of the intermodulation distortion in the second-order path is relatively small.
Large distortion of nonlinear modulation device at high frequency
Phase that occurs at a relatively low frequency
Non linear in a relatively low frequency the magnitude of the intermodulation distortion
Is substantially equal to the magnitude of the distortion of the linear modulation device.
The intermodulation distortion that occurs at higher frequencies
Adjust tilt without substantially changing the size
2. The method according to claim 1, further comprising the steps of:
Law. 10. A frequency function in each secondary path.
Adjusting means for adjusting the relative amplitude of the signal in the secondary path
The stage further compares the magnitude of the intermodulation products in each secondary path.
Realizes distortion of nonlinear devices at relatively low frequencies
Means of qualitative equalization, and the ratio that occurs in each secondary path
Relatively high intermodulation products at relatively high frequencies
Magnitude of nonlinear modulation device distortion at high frequencies
At a relatively low frequency so that
Do not substantially change the magnitude of the intermodulation product
And a means for adjusting the tilt.
A distortion correction circuit according to claim 3. 11. The tilt in each of said secondary paths
Means for adjusting each of said secondary paths separately
Variable resistor, capacitor and variable inductor to be grounded
The strain according to claim 10, wherein the strainer has a cutter.
Correction circuit. 12. A frequency function in each secondary path.
Adjusting means for adjusting the relative amplitude of the signal in the secondary path
The stage further compares the magnitude of the intermodulation products in each secondary path.
Realizes distortion of nonlinear devices at relatively high frequencies
Means of qualitative equalization, and the ratio that occurs in each secondary path
Relatively low magnitude of intermodulation products at relatively low frequencies
Magnitude of nonlinear modulation device distortion at high frequencies
At a relatively high frequency so that
Do not substantially change the magnitude of the intermodulation product
And a means for adjusting the tilt.
A distortion correction circuit according to claim 3. 13. The tilt in each of said secondary paths
Means for adjusting each of said secondary paths separately
Variable resistor, capacitor and variable inductor to be grounded
13. The strain according to claim 12, wherein the strainer has a cutter.
Correction circuit.